Single Node vs High Availability
Choose single-node (evaluation) or HA (production) topology based on uptime, failover, and scaling requirements.
After choosing your installation method, decide whether to run a single node or high availability (HA).
When To Use This Page
Use this page after choosing an installation method and before production sizing or infrastructure provisioning.
flowchart TD
subgraph flow["Single Node vs High Availability"]
A[Decision Maker] -->|Deployment Topology| B{Production?}
B -->|Yes| C[High Availability<br />Recommended]
B -->|No| D[Single Node<br />Evaluation / Dev]
end
Comparison
| Single Node | High Availability | |
|---|---|---|
| Use case | Evaluation, dev, small teams | Production, mission-critical |
| Nodes | 1 | 2+ (3+ recommended) |
| Database | Bundled or external | External PostgreSQL required |
| Storage | Local PVC | Shared object storage (S3, GCS, Azure Blob) |
| Failover | None — manual recovery | Automatic |
| Downtime for maintenance | Yes | No — rolling updates |
| Horizontal scaling | No | Yes |
HA Requirements
To run High Availability, you need:
- External database — Managed PostgreSQL (AWS RDS, Azure Database for PostgreSQL, or Google Cloud SQL)
- Shared object storage — S3, GCS, or Azure Blob so all nodes share binaries
- Load balancer — Ingress (Kubernetes) or Route (OpenShift) with health checks
- Multiple replicas —
artifactory.artifactory.replicaCountof 2+ (3+ recommended) - Shared keys — Master Key and Join Key in Kubernetes Secrets
flowchart TD
subgraph arch["High Availability Architecture"]
LB[Load Balancer]
LB --> N1[Node 1]
LB --> N2[Node 2]
LB --> N3[Node 3]
N1 & N2 & N3 --> PG[(PostgreSQL)]
N1 & N2 & N3 --> OS[Object Storage]
end
Migration Path
You can start with a single-node deployment and migrate to HA later by adding an external database, shared storage, and additional replicas. See High Availability Planning.
Practical Decision Thresholds
Use these thresholds as a quick decision aid:
| Signal | Start with Single Node | Start with HA |
|---|---|---|
| Availability target | Non-critical workloads | Strict uptime/SLA targets |
| Team size / usage | Small team, low concurrency | Multiple teams, sustained parallel CI/CD usage |
| Operational model | Evaluation or short-lived environments | Long-lived production with planned maintenance windows |
| Recovery tolerance | Manual recovery acceptable | Automatic failover required |